Dataset · opendiffusionai
Selection of 130 thousand photos from Pexels
opendiffusionai/pexels-photos-janpf
A set of aesthetic photos from the Pexels stock photo site for training generative models and classifiers.
What's inside
Inside, about 130,000 images taken from the Pexels stock website based on a link list from a repository for evaluating photo aesthetics are collected. File names are formed based on the MD5 hash of each picture.
What you can do with it
- Training and fine-tuning image generation models (text-to-image)
- Training image classifiers and visual aesthetics assessment models
- Tasks
- Text → image, Image classification
- Size
- 100K–1M rows
- Downloads this month
- 5.6K
- Likes
- 46
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("opendiffusionai/pexels-photos-janpf")